Photorefraction and Optical Damage Resistance Enhancement in Uranium-Doped Lithium Niobate Crystals by Hafnium Co-Doping

Abstract

A series of Hf co-doped uranium-doped lithium niobate (LN:U,Hf) crystals with a diameter of one inch were grown by the modified Bridgman method. XPS analysis showed that U ions coexist in mixed valence states of U⁴⁺, U⁵⁺, and U⁶⁺. At 442 nm, LN:U,Hf_1.0 exhibited a fast photorefractive response of 0.32 s together with a high saturation diffraction efficiency of 82.01%. With increasing Hf concentration, the optical damage resistance was significantly enhanced, and LN:U,Hf_5.0 achieved an optical damage threshold of 2.8 × 10⁵ W/cm². Two-beam coupling experiments indicated that electrons are the dominant charge carriers and diffusion is the main transport mechanism. It demonstrates that co-doping Hf⁴⁺ provides an effective route to simultaneously enhance photorefractive response and optical damage resistance in LN:U, offering potential for high-power and fast-response photonic devices.

1. Introduction

Lithium niobate (LiNbO₃, LN), widely recognized as the “silicon of photonics,” is a multifunctional crystal that combines excellent acousto-optic, electro-optic, piezoelectric, nonlinear-optical, and photorefractive properties [1,2,3,4]. Owing to this unique combination of physical properties, LN has long played an important role in nonlinear frequency conversion, electro-optic modulation, holographic storage, and integrated photonics [5,6,7,8]. With the rapid development of information-intensive technologies, such as artificial intelligence, lithium niobate on insulator (LNOI) has attracted growing attention as a promising platform for compact and high-speed optical information processing. The photorefractive (PR) effect discovered in the LN single crystal has recently been of great value in LN-based photonic devices. For example, a previous study has shown that LN micro resonators possess markedly fast PR response for dynamic device operation [9]. In addition, PR induced intracavity Bragg scattering has been observed in LN ring resonators, indicating that PR driven refractive-index modulation can directly alter cavity-mode coupling and device behavior [10]. It is well known that LN single crystals are the substrate materials for LNOI thin films. Therefore, controlling the PR properties of LN single crystals, particularly the response speed, is of practical significance for on-chip photonic devices.

Doping is one of the most effective strategies for tailoring the PR properties of LN, primarily through modulation of the defect structure. Early studies mainly focused on introducing low valence PR ions, such as Fe, Cu, Ce, and Mn, into the Li-site to adjust the PR characteristics. For example, LN: Fe exhibits a response time of 32 s at 488 nm, which is about four times faster than that of LN (178 s) while maintaining a high saturated diffraction efficiency of 65% [4]. However, the corresponding response time remained relatively long. In recent years, it has been shown that doping high valence PR ions, such as V and Mo, into the Nb-site can significantly enhance photorefraction in LN crystals [11,12,13]. For example, LN: Mo exhibited shortened response times of 5.5 s and 7 s at 488 nm and 532 nm, respectively, whereas the corresponding saturation diffraction efficiencies were 28.5% and 18.2% [13]. LN: V achieved a response time as short as 0.57 s in the visible region, yet the saturation diffraction efficiency was only 1.1% [11,12]. Recently, our group found that uranium (U) with high valence was a novel PR dopant [14]. LN: U possessed a fast response time of 1.98 s and a saturation diffraction efficiency of 67.15%.

For PR ions doped LN, co-doping optical damage resistant (ODR) ions is beneficial for enhancing PR response speed or improving its resistance to optical damage because they can get rid of intrinsic defects in LN [15,16,17,18,19,20]. In LN, ODR dopants mainly include divalent ions such as Mg²⁺, trivalent ions such as In³⁺, and tetravalent ions such as Hf⁴⁺ [21,22,23]. Following this strategy, our group introduced Mg²⁺ and In³⁺ into LN: U crystals. LN: U,Mg crystals exhibited an optical damage threshold of 2.36 × 10⁵ W/cm² at 532 nm, but only maintained a response time comparable to that of LN: U [24]. LN: U,In crystals displayed a saturation diffraction efficiency similar to that of LN: U at 488 nm, but its response time was approximately twice as long [25]. These results indicate that divalent and trivalent ODR ions have not yet achieved the simultaneous enhancement of optical damage resistance and further acceleration of the response speed of LN: U. Therefore, it is necessary to further explore the role of tetravalent ODR ions in LN: U crystals.

Previous studies have shown that doping Hf⁴⁺ can markedly enhance optical damage resistance, reaching a level comparable to that of LN: Mg [23,26,27]. In addition, LN: Mo,Hf crystals show an optical damage threshold three orders of magnitude higher than that of LN: Mo, while maintaining fast response times in the visible range [28]. These findings indicate that co-doping Hf⁴⁺ may be an effective strategy for simultaneously improving PR properties and ODR in LN: U. In this work, a series of Hf⁴⁺ co-doped LN: U crystals were grown by the modified Bridgman method. The crystal growth, PR properties, and ODR were investigated.

2. Materials and Methods

2.1. Crystal Growth and Sample Preparation

When high valence photorefractive ions (such as U⁶⁺) are incorporated into the LN lattice, they preferentially occupy the Nb-site, leading to the formation of additional anti-site defects Nb_Li⁴⁺. To eliminate these defect centers, a sufficiently high concentration of ODR ions must be co-doped to compensate for the charge imbalance and reduce lattice disorder. Based on our previous work, the UO₂ concentration was fixed at 0.6 mol% [14]. According to the lithium-vacancy model, the HfO₂ concentration required for defect elimination can be estimated as follows. First, the intrinsic anti-site defects Nb_Li⁴⁺ in congruent CLN require approximately 1.28 mol% HfO₂ for complete compensation [26,27]. Second, according to Equation (1), the incorporation of 0.6 mol% U ions is expected to introduce an additional ~3 mol% Nb_Li⁴⁺ anti-site defects. For the convenience of estimation, it is assumed here that all incorporated U ions are oxidized to U⁶⁺ to obtain the maximum estimated threshold concentration. According to Equation (2), the removal of these U-induced defects requires approximately 1.2 mol% HfO₂. Therefore, it is calculated that an HfO₂ concentration of approximately 2.48 mol% is required to eliminate all Nb_Li⁴⁺ anti-site defects in the LN: U system. Considering that the reported threshold concentration for mono-doped Hf⁴⁺ in LN is in the range of 2.0–4.0 mol% [26,27], the estimated threshold concentration of Hf⁴⁺ in LN: U is 4.48–6.48 mol%. It should be emphasized that this value is only an approximate estimate based on a simplified charge compensation model for threshold range analysis, rather than an exact experimental threshold. Accordingly, Hf⁴⁺ concentrations of 1.0, 5.0, and 6.0 mol% were chosen in this work, and the obtained crystals were denoted as LN: U,Hf_1.0, LN: U,Hf_5.0, and LN: U,Hf_6.0.

$$8\mathrm{L}\mathrm{i}\mathrm{N}\mathrm{b}{\mathrm{O}}_3+2\mathrm{H}\mathrm{f}{\mathrm{O}}_2+\mathrm{U}{\mathrm{O}}_3=\left({\mathrm{V}}_{\mathrm{Li}}^{-}\right)\left({\mathrm{U}}_{\mathrm{Nb}}^{+}\right){\mathrm{O}}_3+8\left({\mathrm{Hf}}_{\mathrm{Li}}^3{)}_{1/4}\right({\mathrm{V}}_{\mathrm{Li}}^{-}{)}_{3/4}\left(\mathrm{N}\mathrm{b}{\mathrm{O}}_3\right)+3{\mathrm{L}\mathrm{i}}_2\mathrm{O}$$

(1)

$$2\mathrm{H}\mathrm{f}{\mathrm{O}}_2+5\left({\mathrm{Nb}}_{\mathrm{Li}}^4{)}_{1/5}\right({\mathrm{V}}_{\mathrm{Li}}^{-}{)}_{4/5}\left(\mathrm{N}\mathrm{b}{\mathrm{O}}_3\right)+2\mathrm{L}\mathrm{i}\mathrm{N}\mathrm{b}{\mathrm{O}}_3=8\left({\mathrm{Hf}}_{\mathrm{Li}}^3{)}_{1/4}\right({\mathrm{V}}_{\mathrm{Li}}^{-}{)}_{3/4}\left(\mathrm{N}\mathrm{b}{\mathrm{O}}_3\right)+{\mathrm{L}\mathrm{i}}_2\mathrm{O}$$

(2)

The raw materials were Li₂CO₃ (99.99%, Shanghai Titan Scientific Co., Ltd., Shanghai, China), Nb₂O₅ (99.99%, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China), HfO₂ (99.99%, Meryer (Shanghai) Chemical Technology Co., Ltd., Shanghai, China), and depleted UO₂ (99.99%, CNNC North Nuclear Fuel Element Co., Ltd., Baotou, China). Among them, Nb₂O₅ was pre-calcined at 800 °C for 5 h to remove fluorine impurities. The accurately weighed powders were thoroughly mixed in an agate mortar and then placed in a corundum crucible. The mixture was sintered at 1150 °C for 7 h in a high-temperature muffle furnace to obtain the polycrystalline material. Subsequently, the polycrystalline material was loaded into a platinum crucible and heated to 1320 °C in the crystal-growth furnace to ensure complete melting. After seed-crystal insertion, the automatic lowering program was initiated, and the crystal-growth period lasted approximately 20 days. Upon completion of growth, the crystals were cooled at a controlled rate of 40 °C/h. Finally, electrical poling was performed at 1200 °C to obtain a single-domain state. The y-cut crystals with thicknesses of 1 mm and 3 mm were obtained after grinding and polishing to optical grade for measurements.

2.2. Property Characterization

The actual concentrations of U and Hf in the crystals were determined by inductively coupled plasma atomic emission spectrometry (ICP-OES), and the effective segregation coefficient (k_eff) was calculated using the following formula:

$$k_{\mathit{eff}}={\displaystyle \frac{C_s}{C_1}}$$

(3)

where Cs represents the actual ion concentration in the crystal, and C₁ represents the initial ion concentration in the melt.

Crystal growth was carried out using a self-designed modified crucible-lowering furnace, and the temperature was controlled by a VBF1600 high-temperature vertical lowering furnace control system manufactured by Shanghai Jingcui Materials Technology Co., Ltd. (Shanghai, China). The elemental composition of the samples was analyzed using a ZEISS GeminiSEM 300 microscope (Zeiss, Oberkochen, Germany). Powder X-ray diffraction (XRD) patterns were collected on a Dandong Tongda TD3500 diffractometer (Dandong Tongda Technology Co., Ltd., Dandong, China) with Cu Kα radiation (λ = 0.15406 nm), over a 2θ range of 5–120° with a step size of 0.02°. Rocking curves were measured using a SmartLab X-ray diffractometer (Rigaku, Tokyo, Japan). X-ray photoelectron spectroscopy was performed using an XPS spectrometer (K-Alpha+, Thermo Fisher Scientific, i.e., Waltham, MA, USA) with an Al Kα microfocus monochromatic source over a range of 30–400 μm and a step size of 5 μm. The segregation coefficient was determined by inductively coupled plasma optical emission spectrometry, ICP-OES (Optima 8000, PerkinElmer, i.e., Waltham, MA, USA). The Cs values used for calculating the effective segregation coefficient were obtained from samples collected from the constant-diameter region near the shoulder of the crystals. The OH⁻ absorption spectra were recorded using a Nicolet 10 Fourier-transform infrared spectrometer (Thermo Scientific, i.e., Madison, WI, USA) in the range of 1900–3900 cm⁻¹.

The PR properties of LN: U,Hf crystals were investigated by two-wave coupling experiments using 3 mm thick samples. The experimental optical setup is shown in Figure 1. The system consisted of a beam splitter (BS), mirrors (M₁ and M₂), shutters (S₁ and S₂), and detectors (D₁ and D₂). Laser sources with wavelengths of 442 nm, 532 nm, and 671 nm were employed. The single-beam intensity was 400 mW/cm² for the 442 nm and 532 nm lasers and 3000 mW/cm² for the 671 nm laser. The two coherent beams were incident on the sample at an angle of 15°, forming an interference grating inside the crystal. The variation in diffraction efficiency was recorded in real time by the detectors.

The diffraction efficiency (η) of a crystal is defined as the ratio of the diffracted intensity (I_d) to the total transmitted intensity (I_d + I_t):

$$\eta ={\displaystyle \frac{I_d}{I_d+I_t}}\times 100\%$$

(4)

The temporal evolution of the diffraction efficiency, η(t), can be fitted by the following function, from which the saturation diffraction efficiency (η_s) and response time (τ_r) are obtained:

$$\eta \left(t\right)={\eta }_s[1-exp(-t/{\tau }_r){]}^2$$

(5)

The expressions for the change in refractive index (Δn) and photorefractive sensitivity (S) are given by

$$∆n={\displaystyle \frac{\lambda \cos \theta }{\pi d}}\arcsin \sqrt{\eta s}$$

(6)

$$S={\displaystyle \frac{1}{I_d}}{{\displaystyle \frac{d\sqrt{\eta }}{\mathit{dt}}}|}_{t=0}$$

(7)

where λ denotes the laser wavelength, d signifies the crystal thickness, θ represents the internal half-angle between the writing beams, and I corresponds to the total incident light intensity.

The optical damage resistance of the crystals was evaluated by the transmitted-beam distortion method, as schematically illustrated in Figure 2. The incident laser beam was focused by a convex lens with a focal length of f = 100 mm and then normally irradiated onto the crystal sample. The optical damage behavior was assessed by monitoring the evolution of the transmitted beam profile on a far-field screen. During the measurement, the laser power was gradually increased until a noticeable distortion of the transmitted spot appeared. The corresponding incident power was recorded, and the optical damage threshold was determined by calculating the associated power density at the focal spot.

3. Results and Discussions

3.1. Crystal Structure and Quality

Using the modified Bridgman method, one-inch LN: U,Hf crystals with Hf⁴⁺ concentrations of 1.0 and 5.0 mol% and lengths of approximately 7–8 cm were successfully grown. As shown in Figure 3a, both crystals exhibit a reddish-brown color. However, bulk growth of LN: U,Hf_6.0 was unsuccessful, and only a tiny crystal grain was obtained, which was insufficient for characterization. As shown in Figure 3b, all diffraction peaks match the standard LiNbO₃ pattern (PDF#74-2238), and no secondary phase is detected. The high-resolution X-ray rocking curves in Figure 3c,d show that the full widths at half maximum of the (006) reflections are 2.5146′ and 3.1176′ for LN: U,Hf_1.0 and LN: U,Hf_5.0, respectively. Their crystallization quality is superior to that of LN: U, In_6.0 crystals [25], although inferior to that of LN: U_0.6 crystals [14], which may be attributed to the lattice distortion and internal stress caused by co-doping Hf⁴⁺.

The effective segregation coefficients calculated from the ICP-OES measurements are summarized in

Table 1. The effective segregation coefficients.

Sample	LN: U,Hf1.0		LN: U,Hf5.0		LN: U,Hf6.0
Ion	Hf	U	Hf	U	Hf	U
Effective segregation coefficient (keff)	1.2184	0.8267	1.1647	0.7367	1.3258	0.485

. In crystal growth, a segregation coefficient closer to 1 indicates that the dopant ions can be incorporated into the lattice more easily, thereby promoting compositional uniformity and stable growth. The results reveal that, with increasing Hf concentration, the effective segregation coefficient of U gradually decreases and deviates further from 1. This trend increases compositional instability during crystal growth, thereby making crystal growth more difficult. The segregation coefficients of the constituent elements in LN: U,Hf_6.0 deviated significantly from 1, which might be the reason for the failure of crystal growth of this crystal.

To further investigate the crystal structure, the XRD data of LN:U,Hf_1.0 were refined by the General Structure Analysis System (GSAS), and the refinement profile is shown in Figure 4. The refinement converged with reliability factors of R_p = 5.94% and R_wp = 3.44%, with a goodness-of-fit (Gof) value of 3.01, indicating a reasonable agreement between the calculated and observed patterns. The refined crystallographic parameters are summarized in

Table 2. Rietveld refinement data for LN: U,Hf_1.0.

Formula	LN: U0.6, Hf1.0
Space group	R3c
a (Å)	5.15214(4)
b (Å)	5.15214
c (Å)	13.86347(10)
A (°)	90
b (°)	90
γ (°)	120
V (Å3)	318.6972(27)
Rp (%)	5.94%
Rwp (%)	3.44%
Gof	3.01

. LN:U,Hf_1.0 crystallizes in the R3c space group, with lattice parameters of a = b = 5.15214(4) Å and c = 13.86347(10) Å, corresponding to a unit-cell volume of 318.6972(27) ų. The atomic coordinates and occupancies are listed in

Table 3. Fractional atomic coordinates and equivalent isotropic displacement parameters for LN: U,Hf_1.0.

LN: U0.6, Hf1.0	Site	X	Z	Ueq	Occupancy
Li1	Li	0.00000	0.28020	0.0147	0.9778
Nb1	Nb	0.00000	0.00000	0.0072	0.9815
O	O	0.04419	0.06271	0.0104	1.0000
Hf1	Li	0.00000	0.28020	0.0147	0.0100
Nb2	Li	0.00000	0.28020	0.0147	0.0122
U1	Nb	0.00000	0.00000	0.0072	0.0064

. It can be seen that U ions occupy the normal Nb-site with an occupancy of 0.0064, whereas Hf ions mainly enter the Li-site with an occupancy of 0.0100. In addition, anti-site defects Nb_Li⁴⁺ are still present at the Li-site with an occupancy of 0.0122.

3.2. Analysis of Ionic Valence States

The valence states of U ions in the crystals were analyzed by XPS. As shown in Figure 5, the U4f XPS spectrum of the LN: U,Hf crystal can be deconvoluted into six characteristic peaks. According to previous reports, the U4f core level is split into two components, U 4f_7/2 and U 4f_5/2, due to spin–orbit coupling, with a characteristic binding-energy separation of approximately 10–12 eV [29,30]. In the present measurements, the energy difference between the two peak groups (Peak 1(379.8 ev), Peak 2(378.8 ev), Peak 3(382.4 ev)) and (Peak 4(388.3 ev), Peak 5(390.8 ev), Peak 6(393.7 ev)) is consistent with this splitting behavior. Accordingly, the six fitted peaks were assigned to three uranium valence states: Peaks 1 and 4 to U⁴⁺, Peaks 2 and 5 to U⁵⁺, and Peaks 3 and 6 to U⁶⁺. These assignments are consistent with our previous observations in LN: U, LN: U,Mg, and LN: U,In crystals [14,24,25], indicating that U ions coexist in mixed valence states (U⁴⁺, U⁵⁺, and U⁶⁺) in LN: U,Hf.

3.3. Photorefractive Properties

The typical PR writing curves of LN: U,Hf_1.0 and LN: U,Hf_5.0 crystals at different wavelengths are shown in Figure 6, and the corresponding performance parameters are summarized in Figure 7. The LN: U,Hf series exhibits a clear wavelength dependence. At 442 nm, the diffraction efficiency, sensitivity, and refractive index modulation reach their maximum values, while the response time is the shortest. As the wavelength increases to 532 nm and 671 nm, the overall PR performance gradually decreases. Both the refractive index change and sensitivity show pronounced wavelength dependence. LN: U,Hf_1.0 maintains a relatively large Δn over 442, 532, and 671 nm and exhibits significantly higher sensitivity than LN: U,Hf_5.0, especially at 442 nm. By contrast, increasing the Hf⁴⁺ concentration to 5.0 mol% reduces Δn and sensitivity, indicating a weakened PR response. This suggests that co-doping moderate Hf⁴⁺ is more favorable for preserving the PR effect. At 532 nm, the response times of both Hf co-doped samples became noticeably longer, which is likely related to the reduced optical absorption at this wavelength and the consequently lower excitation efficiency of photogenerated charge carriers. At the longer wavelength of 671 nm, both samples exhibited response times exceeding 240 s. However, their saturation diffraction efficiencies were still more than three times that of LN: U_0.6.

Notably, the LN: U,Hf_1.0 crystal exhibited an exceptionally short response time of 0.32 s at 442 nm, which is approximately six times faster than that of LN: U_0.6 (1.98 s) [14], nearly three times faster than LN: Mo,Hf (0.9 s) [28], and about nineteen times faster than LN: U,In_2.0 (6.14 s) [25]. Meanwhile, it maintained a high saturation diffraction efficiency of 82.01%, significantly higher than those of LN: U_0.6 (50%) [14] and LN: Mo,Hf_3.5 (46.07%) [28]. LN: U,Hf_5.0 exhibited a response time of 1.14 s at 442 nm. Although slower than that of LN: U,Hf_1.0, it remained faster than LN: U_0.6. These results clearly indicate that appropriate co-doping with Hf⁴⁺ can substantially accelerate the PR response of LN: U crystals at 442 nm.

The dominant photogenerated charge carriers and their migration mechanism were examined using two-beam energy coupling experiments. As shown in Figure 8, for the LN: U,Hf_1.0 crystal, energy is transferred from the signal beam (S beam) to the reference beam (R beam) along the -C axis, indicating that electrons are the dominant charge carriers. The observed energy transfer behavior further suggests that diffusion is the primary migration mechanism [25]. A similar energy transfer behavior was also observed in the LN: U,Hf_5.0 crystal.

3.4. Optical Damage Resistance

The optical damage resistance of the crystals was evaluated at 442 nm using the transmitted-beam distortion method. As shown in Figure 9, after irradiation with a laser intensity of 2.80 × 10⁵ W/cm² for 5 min, the transmitted beam of the LN: U,Hf_1.0 crystal exhibited severe distortion (Figure 9b), whereas the LN: U,Hf_5.0 sample maintained a circular beam profile without noticeable distortion (Figure 9d). Further measurements revealed that the optical-damage threshold of LN: U,Hf_1.0 was approximately 6.0 × 10² W/cm², while that of LN: U,Hf_5.0 reached 2.8 × 10⁵ W/cm². This value is not only much higher than that of the LN: U,Hf_1.0, but also higher than LN: Mo,Hf [28] and comparable to LN: U,Mg_7.0 [24]. These results clearly demonstrate that high-concentration Hf co-doping can markedly enhance the ODR of LN: U.

3.5. OH⁻ Absorption and Raman Spectroscopic Analysis

FTIR spectroscopy was used to determine whether the concentration of ODR ions had reached the threshold level by monitoring the position of the OH⁻ stretching vibration band. For congruent LN, the OH⁻ absorption peak is typically located near 3484 cm⁻¹ and is associated with defect-related OH⁻ complexes [31,32]. When optical damage-resistant ions are incorporated into the lattice and get rid of anti-site defects Nb_Li⁴⁺, the OH⁻ absorption band shifts toward higher wavenumbers [27,32]. As shown in Figure 10, the OH⁻ absorption peak of LN: U,Hf_1.0 was located at 3477 cm⁻¹, which is close to that of CLN, indicating that the 1 mol% had not yet reached the concentration threshold of Hf⁴⁺ in LN: U,Hf. In contrast, the OH⁻ absorption band of LN: U,Hf_5.0 exhibited a pronounced blue shift to 3621 cm⁻¹. This shift was attributed to the formation of Hf-related OH complexes, suggesting that 5 mol% exceeded the concentration threshold of Hf⁴⁺ in LN: U,Hf [27], and all Nb_Li⁴⁺ anti-site defects in LN: U were eliminated. These observations are consistent with the theoretical estimates according to Equations (1) and (2), and also provide a reasonable explanation for the significantly enhanced ODR observed in LN: U,Hf _5.0.

Raman spectroscopy further corroborated the above conclusions. As shown in Figure 11, the Raman spectra of the LN:U,Hf are generally similar to that of CLN, indicating that co-doping Hf⁴⁺ does not alter the basic lattice framework of lithium niobate. Since the broad band in the 550–650 cm⁻¹ region is likely composed of overlapping modes, Gaussian fitting was performed to separate the individual components and determine their peak positions more accurately. The main Raman bands are located near 152, 237, 255, and 618 cm⁻¹. Among them, the band near 255 cm⁻¹ is assigned to the Nb_Li–O stretching vibration associated with Nb_Li⁴⁺ defects. The intensity of this band is stronger in LN: U,Hf_1.0 than in the other samples, indicating that a certain amount of Nb_Li⁴⁺ defects still remains at co-doping Hf⁴⁺ with low concentration. With increasing Hf⁴⁺ concentration, this band weakens, suggesting that more Nb_Li⁴⁺ defects are cleared and more Nb⁵⁺ are repelled back to their regular lattice sites. As shown in

Table 4. Fitted peak positions and FWHM values of the Raman bands in the 550–650 cm⁻¹ region.

Sample	Raman Band/cm−1	FWHM/cm−1
LN:U,Hf1.0	576.46	30.92
	616.85	51.63
LN:U,Hf5.0	574.32	34.16
	616.06	57.22

, the broad band in the 550–650 cm⁻¹ region can be deconvoluted into two peaks at approximately 575–576 cm⁻¹ and 616–618 cm⁻¹, corresponding to O–O and Nb–O stretching vibrations, respectively. With increasing Hf concentration, only slight changes in peak position are observed, whereas the FWHM values increase slightly. This indicates that Hf co-doping does not significantly alter the basic lattice structure of LN, but may increase local structural disorder, thereby leading to enhanced local lattice distortion.

3.6. Absorption Spectra and Discussion of Photorefractive Centers

Figure 12a shows the optical absorption spectra of the different samples. According to the Li-vacancy model, the absorption edge is generally defined as the wavelength corresponding to an absorption coefficient of 20 cm⁻¹ [33]. According to this criterion, the absorption edges of CLN, LN:U_0.6, LN:U,Hf_1.0, and LN:U,Hf_5.0 were determined to be 318, 528, 420, and 415 nm, respectively. Obviously, compared with CLN, all doped crystals exhibit an evident red shift in the absorption edge. For LN, the position of the absorption edge is highly sensitive to the defect structure. It has been reported that the absorption edge is related to the transition energy from the O2p states to the Nb4d states, and can be strongly affected by intrinsic defects and defect-related local electronic states [33,34,35,36]. In particular, Li-site related disorder and cation mixing can modify the local electronic structure near the band edge, thereby causing red or blue shifts in the absorption edge [36]. According to our former work, U^5+/6+ occupies the Nb-site and pushes the Nb ions to the Li-site, which aggravates the Li/Nb cation mixing. Thus, the band gap of LN: U_0.6 was narrower than CLN, which makes a strong redshift of the absorption edge [14]. In contrast, relative to LN: U_0.6, the LN: U,Hf crystals show a blue shift. This behavior can be reasonably attributed to the preferential occupation of the Li-site by Hf⁴⁺ ions, and repels anti-site defects Nb_Li⁴⁺ back to the regular Nb-site, reduces defect-related localized states, and thus widens the effective band gap [26,27,33]. As the Hf concentration increased from 1.0 mol% to 5.0 mol% and exceeded the threshold level, the absorption edge underwent a further slight blue shift, consistent with the continued evolution of the defect structure.

As can be seen from Figure 12b, compared with CLN, the LN:U,Hf co-doped crystals exhibit much stronger absorption in the wavelength range of about 400–600 nm. The transmittance of LN: U,Hf_1.0 is about 30% at around 450 nm and 39% near 520 nm, and then rises to about 61% in the 600–800 nm region; for LN: U,Hf_5.0, the corresponding values are about 35%, 46%, and 69%, respectively. Such a broad absorption band indicates the presence of multiple PR active centers that can be excited by different wavelengths to generate carriers, which also explains why the LN: U,Hf crystals exhibit enhanced PR performance at 442, 532, and 671 nm. In particular, U ions with three valence states coexist in LN: U,Hf, and previous studies on LN: U and LN: U,In have shown that mixed valence U ions related defect clusters can serve as PR centers [14,25]. In addition to these U-related centers, intrinsic defect states, possibly including oxygen-related defects, may also participate in photocarrier generation, trapping, and transport [34,35]. Therefore, the spectral evolution shown in Figure 12 provides indirect but useful evidence for understanding the correlation between defect structure and PR performance. Nevertheless, the exact nature and specific roles of these PR centers still require further investigation by complementary characterization techniques.

4. Conclusions

Hf⁴⁺ co-doped LN: U crystals, one inch in diameter, were grown by the modified Bridgman method. LN: U,Hf_1.0 exhibited a fast response time of 0.32 s and a high saturation diffraction efficiency of 82.01% at 442 nm. Its response time is approximately six times faster than that of LN: U and about nineteen times faster than LN: U,In. When the doping concentration of Hf⁴⁺ reached 5 mol%, the optical damage threshold of LN: U,Hf was as high as 2.8 × 10⁵ W/cm². The result of the OH⁻ spectrum showed that 5 mol% exceeded the real threshold concentration of Hf⁴⁺ in LN: U. Two-beam coupling measurements showed that electrons are the dominant carriers and diffusion is the primary transport mechanism. The defects were also discussed based on the Li-vacancy model. Overall, Hf co-doping enables both fast PR response and high optical damage resistance in LN: U crystals, making this system promising for high-power, fast-response photonic devices.